When trying to add a node to a cluster, the following error occurs with the status of the node showing pending in the UCP UI:
This can also manifest itself from the CLI with the following node state:
Awaiting healthy status in classic node inventory
Most likely what has happened is the Engine IDs have been duplicated, prior to joining the UCP Swarm cluster. In most cases, this is a consequence of cloning VMs or using AWS snapshots with Engine already installed.
The Engine ID is generated on Docker's first run and stored in
When creating a VM template or AWS snapshot, make sure
/etc/docker/key.json is not included in the image.
Make sure your cluster is healthy and all your managers are working correctly.
To recover the existing installation, remove
/etc/docker/key.json and restart Docker daemon on the impacted node.
Then restart the Docker daemon on the Swarm Leader node.
More details on node status codes can be found on docs.docker.com.